Abstract
Abstract
Objectives
To examine the association between baseline sleep quality and subsequent glycaemic outcomes in adults with type 2 diabetes mellitus (T2DM), using repeated fasting blood glucose measurements and longitudinal glycaemic pattern analyses.
Design
Prospective cohort study.
Setting
The Taizhou Diabetes Family-based Cohort in Zhejiang Province, China.
Participants
A total of 1421 adults with T2DM with baseline sleep quality assessment and follow-up fasting blood glucose measurements were included.
Main outcome measures
Baseline sleep quality was assessed using the Pittsburgh Sleep Quality Index (PSQI). Follow-up fasting blood glucose measurements were obtained through 30 June 2025. The primary outcomes were continuous fasting blood glucose levels and binary fasting blood glucose status defined as ≥7.0 mmol/L during follow-up. Secondary outcomes included glycaemic variability and latent glycaemic trajectory classes.
Results
Poorer baseline sleep quality was associated with less favourable glycaemic outcomes during follow-up. In the fully adjusted model, each 1-point increase in PSQI score was associated with a 0.096 mmol/L higher fasting blood glucose level (95% CI 0.081 to 0.111), and poor sleep quality was associated with a 0.669 mmol/L higher fasting blood glucose level (95% CI 0.547 to 0.791). For the binary outcome, each 1-point increase in PSQI score was associated with higher odds of fasting blood glucose ≥7.0 mmol/L (OR 1.10, 95% CI 1.08 to 1.12), and poor sleep quality was associated with 1.61-fold higher odds of fasting blood glucose ≥7.0 mmol/L (95% CI 1.37 to 1.91). Higher PSQI scores were also associated with greater glycaemic variability and higher odds of membership in less favourable glycaemic trajectory groups.
Conclusions
Among adults with T2DM, poorer baseline sleep quality was associated with higher fasting blood glucose levels, poorer glycaemic control, greater glycaemic variability and less favourable long-term glycaemic trajectories. These findings suggest that sleep quality may be a relevant and easily assessable marker for identifying individuals at risk of suboptimal long-term glycaemic management.
Keywords: Glycemic Control; Diabetes Mellitus, Type 2; EPIDEMIOLOGY
STRENGTHS AND LIMITATIONS OF THIS STUDY.
This study used prospective cohort data from adults with type 2 diabetes mellitus, allowing temporal assessment of baseline sleep quality in relation to subsequent glycaemic outcomes.
Repeated fasting blood glucose measurements during follow-up enabled evaluation of average glycaemic levels, binary glycaemic control, glycaemic variability and longitudinal trajectory patterns.
Multiple analytical approaches, including mixed-effects models, sensitivity analyses and latent class trajectory modelling, were used to assess the robustness and heterogeneity of the associations.
Sleep quality was assessed only at baseline, and changes in sleep quality during follow-up could not be evaluated.
Long-term glycaemic patterns were primarily based on repeated fasting blood glucose rather than repeated glycated haemoglobin (HbA1c) measurements, which may not fully capture longer-term glycaemic exposure.
Introduction
Type 2 diabetes mellitus (T2DM) has become a major public health challenge worldwide and is associated with multiple adverse health outcomes and substantial disease burden.1–3 As chronic hyperglycaemia plays a central role in the development and progression of these complications, maintaining favourable glycaemic control remains a key goal of diabetes management.4 5 In clinical practice and public health, considerable attention has traditionally focused on medication use, dietary patterns, physical activity and other established determinants of glycaemic control.3 4 However, increasing evidence suggests that sleep may also play an important role in metabolic regulation and diabetes management.6 7
Sleep disturbance is common among adults with T2DM,8 9 who often report poorer sleep quality, shorter or more fragmented sleep10 and greater daytime dysfunction.11 12 Increasing evidence suggests that sleep may be involved in metabolic regulation, and sleep quality has therefore attracted growing attention as a potential correlate of glycaemic status in this population.8 9 13 Several studies have suggested that poorer subjective sleep quality is associated with worse glycaemic control, including higher fasting glucose and glycated haemoglobin (HbA1c) levels.9 14 These findings indicate that sleep quality may be relevant to glycaemic management in adults with T2DM.4 6
Current evidence on the relationship between sleep quality and glycaemic status in adults with T2DM remains incomplete.6 9 Previous studies have focused mainly on cross-sectional associations and have typically related sleep quality to glycaemic measures assessed at a single time point, such as fasting glucose or HbA1c.15 16 Relatively less is known about whether baseline sleep quality is associated with subsequent glycaemic changes or with longer-term patterns of glycaemic progression over time.14
To extend the current evidence base, this study used prospective data from adults with T2DM and repeated fasting blood glucose measurements during follow-up to examine the associations of baseline sleep quality with subsequent glycaemic outcomes. We also explored individual Pittsburgh Sleep Quality Index (PSQI) components, subgroup differences, glycaemic variability and latent glycaemic trajectories to further characterise the potential relevance of sleep quality to long-term glycaemic regulation.
Methods
Study design and participants
The Taizhou Diabetes Family-based Cohort was established in Taizhou, Zhejiang Province, China. From May to October 2021, individuals with T2DM from six county-level areas of Taizhou were recruited as probands, and their first-degree, second-degree and third-degree relatives were subsequently enrolled. A total of 3974 participants were enrolled at baseline. Eligibility criteria included mental competence, the ability to complete the questionnaire and physical examination, provision of written informed consent and residence in the study area for at least 1 year with no plans to relocate within the subsequent 3 years. In addition, probands were required to have a confirmed diagnosis of diabetes and at least one first-degree relative (parent, sibling or child) with diabetes. Baseline assessments included face-to-face questionnaire interviews conducted by trained local interviewers; physical examinations with duplicate measurements of height, weight, waist circumference and blood pressure; and collection of venous blood samples after an overnight fast of at least 8 hours for measurement of HbA1c, fasting blood glucose, triglycerides, total cholesterol, high-density lipoprotein cholesterol and low-density lipoprotein cholesterol. This study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology guideline.
For the present analysis, we identified an analytic subcohort of 1788 participants with a confirmed diagnosis of diabetes at baseline from the Taizhou Diabetes Family-based Cohort (figure 1). After excluding participants with missing data on follow-up fasting blood glucose, sleep or other key covariates, 1421 individuals were included in the final analysis. The mean follow-up duration was 3.92±0.16 years, during which 52 940 fasting blood glucose measurements were obtained through 30 June 2025. The study was approved by the Ethics Committee of the School of Public Health, Fudan University, and all participants provided written informed consent.
Figure 1. Flowchart of participant selection. T2DM, type 2 diabetes mellitus; FBG, fasting blood glucose.
Assessment of sleep quality and follow-up fasting blood glucose
Baseline sleep quality was assessed using the PSQI, which comprises seven components: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication and daytime dysfunction. The global PSQI score ranges from 0 to 21, with higher scores indicating poorer sleep quality.17 In the main analyses, the PSQI score was modelled as a continuous variable including per 1-SD increase and as a dichotomous variable, with poor sleep quality defined as a PSQI score >5. Exploratory analyses further evaluated the associations of the seven PSQI components with both continuous and binary fasting blood glucose outcomes. Follow-up fasting blood glucose measurements were obtained from the local chronic disease management database, which integrates records from routine health examinations, diabetes follow-up visits, diabetes registry records and hospital laboratory tests.
Assessment of covariates
Baseline diabetes status was defined by any of the following: self-reported physician-diagnosed diabetes, fasting blood glucose ≥7.0 mmol/L, HbA1c ≥6.5% or current use of glucose-lowering medication.3 Diabetes duration was defined as the interval between the recorded date of diabetes diagnosis, obtained from the local electronic health record system and the baseline survey date. HbA1c was categorised as <7.0% or ≥7.0%. Body mass index (BMI) was calculated as weight in kilogrammes divided by height in metres squared (kg/m2). Hypertension was defined as systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg.18 Dyslipidaemia was defined as the presence of any of the following: triglycerides ≥2.26 mmol/L, total cholesterol ≥6.22 mmol/L, low-density lipoprotein cholesterol ≥4.14 mmol/L or high-density lipoprotein cholesterol <1.04 mmol/L.19 Smoking status and alcohol consumption were each classified as never, former or current based on self-reported baseline information. Physical activity was assessed using the WHO Global Physical Activity Questionnaire (GPAQ) and categorised as insufficient, sufficient or high according to the GPAQ analysis guidelines.20
Statistical analysis
Sample size was estimated for the primary repeated-measures analysis. Assuming a minimum detectable increase of 0.05 mmol/L in fasting plasma glucose per 1-point increase in PSQI score, 90% power, a two-sided significance level of 0.05 and 20% incomplete follow-up, at least 954 participants were required. 1421 participants were finally included. To examine the associations of baseline sleep quality with repeated fasting blood glucose measurements during follow-up, mixed-effects models were fitted to account for the hierarchical structure of the data, with repeated measurements nested within individuals and individuals nested within families. Random intercepts were included for both participant ID and family ID to account for within-individual correlation and potential clustering among participants from the same family, respectively. Continuous fasting blood glucose was analysed using linear mixed-effects models, and binary fasting blood glucose defined using a cut-off 7.0 mmol/L was analysed using mixed-effects logistic regression models. Three progressively adjusted models were fitted: model A adjusted for age and sex; model B further adjusted for BMI, smoking status, alcohol consumption and physical activity; model C additionally adjusted for follow-up time, HbA1c, use of glucose-lowering medication, hypertension and dyslipidaemia. Results are presented as β coefficients or ORs with 95% CIs. Exploratory analyses examined the seven PSQI components in relation to both continuous and binary fasting blood glucose outcomes. Sensitivity analyses included restriction to participants with at least three fasting blood glucose measurements, generalised estimating equation models and mixed-effects models with participant-specific random slopes for follow-up time, while retaining random intercepts for both participant ID and family ID. In addition, latent class trajectory modelling was applied to repeated fasting blood glucose measurements to identify distinct longitudinal glycaemic trajectories; models with one to four classes were compared, and a four-class solution was retained based on model fit. All analyses were performed using R V.4.4.2 (R Foundation for Statistical Computing, Vienna, Austria), and statistical significance was defined as a two-sided p<0.05.
Patient and public involvement
Patients or the public were not involved in the design, conduct, reporting or dissemination plans of this study.
Results
Baseline characteristics
Among the 1421 participants, 562 (39.5%) were classified as having better sleep quality and 859 (60.5%) as having poor sleep quality. Compared with those with better sleep quality, participants with poor sleep quality were older and less likely to be male. Smoking status also differed between the two groups. No significant differences were observed in BMI, diabetes duration, baseline HbA1c category, antidiabetic medication use, alcohol consumption, physical activity, hypertension or dyslipidaemia (all p>0.05). Baseline characteristics according to sleep quality are presented in table 1.
Table 1. Baseline characteristics of patients with diabetes according to sleep quality at baseline.
| Variable | Total | Better sleep quality (n=562) |
Poor sleep quality (n=859) |
P value |
|---|---|---|---|---|
| Age, year | 64.73±10.52 | 63.65±10.70 | 65.44±10.35 | 0.002 |
| Male, n (%) | 668 (47.0) | 288 (51.2) | 380 (44.2) | 0.011 |
| BMI ≥24, n (%) | 870 (61.2) | 359 (64.3) | 511 (59.5) | 0.930 |
| Diabetes duration, year | 7.64±8.98 | 7.20±9.41 | 7.93±8.68 | 0.133 |
| HbA1c ≥7.0%, n (%) | 857 (60.3) | 342 (60.8) | 515 (60.0) | 0.481 |
| Antidiabetic medication, n (%) | 455 (32.0) | 188 (33.5) | 267 (31.1) | 0.380 |
| Smoke, n (%) | ||||
| Never | 1003 (70.6) | 378 (67.3) | 625 (72.8) | 0.036 |
| Former | 145 (10.2) | 70 (12.5) | 75 (8.7) | |
| Current | 273 (19.2) | 114 (20.3) | 159 (18.5) | |
| Alcohol, n (%) | ||||
| Never | 999 (70.3) | 384 (68.3) | 615 (71.6) | 0.393 |
| Former | 72 (5.1) | 29 (5.2) | 43 (5.0) | |
| Current | 350 (24.6) | 149 (26.5) | 201 (23.4) | |
| Physical activity, n (%) | ||||
| Insufficient | 746 (52.5) | 313 (55.7) | 433 (50.4) | 0.149 |
| Enough | 49 (3.4) | 18 (3.2) | 31 (3.6) | |
| High | 626 (44.1) | 231 (41.1) | 395 (46.0) | |
| Hypertension, n (%) | 705 (49.6) | 269 (48.1) | 436 (51.5) | 0.231 |
| Dyslipidaemia, n (%) | 949 (67.6) | 380 (68.7) | 569 (66.9) | 0.525 |
BMI, body mass index; HbA1c, glycated haemoglobin.
Associations of baseline sleep quality with follow-up fasting blood glucose
Poorer baseline sleep quality was associated with higher follow-up fasting blood glucose in both continuous and binary outcome analyses, and the findings were generally consistent across models A–C. In the fully adjusted model, each 1-point increase in PSQI score was associated with a 0.096 mmol/L (95% CI 0.081 to 0.111) higher fasting blood glucose level, and per SD increase was associated with a 0.366 mmol/L (95% CI 0.307 to 0.425) higher blood glucose level. Compared with participants with better sleep quality, those with poor sleep quality had a 0.669 mmol/L (95% CI 0.547 to 0.791) higher fasting blood glucose level. In the binary outcome analysis, higher PSQI scores were also associated with higher odds of fasting blood glucose ≥7.0 mmol/L. In model C, the corresponding ORs were 1.10 (95% CI 1.08 to 1.12) per 1-point increase and 1.44 (95% CI 1.33 to 1.56) per 1-SD increase in PSQI score, while poor sleep quality was associated with 1.61-fold higher odds of fasting blood glucose ≥7.0 mmol/L (95% CI 1.37 to 1.91). Sensitivity analyses yielded broadly similar results. Detailed estimates are presented in table 2 and online supplemental table S1.
Table 2. Associations of baseline sleep quality with follow-up fasting blood glucose across progressively adjusted mixed-effects models.
| Variable | β (95% CI) | ||
|---|---|---|---|
| Model A | Model B | Model C | |
| PSQI total score | 0.097 (0.079 to 0.115)* | 0.099 (0.081 to 0.117)* | 0.096 (0.081 to 0.111)* |
| Per SD increase | 0.373 (0.305 to 0.463)* | 0.382 (0.313 to 0.452)* | 0.366 (0.307 to 0.425)* |
| Poor sleep | 0.639 (0.498 to 0.780)* | 0.645 (0.503 to 0.788)* | 0.669 (0.547 to 0.791)* |
| Variable | OR (95% CI) | ||
| Model A | Model B | Model C | |
| PSQI total score | 1.10 (1.08 to 1.13)* | 1.10 (1.08 to 1.13)* | 1.10 (1.08 to 1.12)* |
| Per SD increase | 1.45 (1.33 to 1.58)* | 1.45 (1.33 to 1.58)* | 1.44 (1.33 to 1.56)* |
| Poor sleep quality | 1.55 (1.30 to 1.85)* | 1.55 (1.30 to 1.86)* | 1.61 (1.37 to 1.91)* |
Model A: adjusted for age and sex.
Model B: model A plus BMI, smoke, alcohol and physical activity.
Model C: model B plus follow-up time, HbA1c, antidiabetic medication, hypertension and dyslipidaemia.
p<0.05, p<0.01.
p<0.001.
HbA1c, glycated haemoglobin; PSQI, Pittsburgh Sleep Quality Index.
Exploratory analyses of the seven Pittsburgh Sleep Quality Index components
In exploratory analyses, several PSQI components were associated with both continuous and binary fasting blood glucose outcomes. Poorer subjective sleep quality, longer sleep latency, shorter sleep duration, more severe sleep disturbances and greater daytime dysfunction were generally associated with higher follow-up fasting blood glucose levels and higher odds of fasting blood glucose ≥7.0 mmol/L. Habitual sleep efficiency showed a positive association mainly at the highest severity level, whereas use of sleeping medication showed no consistent pattern. Detailed component-specific results are presented in table 3 and online supplemental table S2.
Table 3. Associations of the seven PSQI component scores with follow-up fasting blood glucose levels in mixed-effects regression analyses.
| Variable | β (95% CI) | ||
|---|---|---|---|
| Mild | Moderate | Severe | |
| Subjective sleep quality | 0.230 (0.018 to 0.442)* | 0.184 (−0.089 to 0.456) | 1.322 (0.897 to 1.747)† |
| Sleep latency | 0.118 (−0.092 to 0.329) | 0.271 (−0.008 to 0.550) | 0.841 (0.460 to 1.223)† |
| Sleep duration | 0.050 (−0.185 to 0.284) | 0.673 (0.258 to 1.087)* | 0.760 (0.547 to 0.974)† |
| Habitual sleep efficiency | 0.179 (−0.196 to 0.554) | 0.116 (−0.400 to 0.631) | 0.669 (0.477 to 0.861)† |
| Sleep disturbances | 0.301 (0.073 to 0.529)* | 0.622 (0.271 to 0.974)† | −0.115 (−1.419 to 1.188) |
| Use of sleeping medication | 1.048 (−0.259 to 2.355) | 0.593 (−0.394 to 1.580) | 0.346 (−0.382 to 1.073) |
| Daytime dysfunction | 0.015 (−0.212 to 0.242) | 0.537 (0.285 to 0.789)† | 0.812 (0.450 to 1.174)† |
| aOR (95% CI) | |||
| Subjective sleep quality | 1.25 (0.94 to 1.66) | 1.29 (0.90 to 1.85) | 4.10 (2.32 to 7.24)† |
| Sleep latency | 1.19 (0.90 to 1.57) | 1.32 (0.92 to 1.91) | 2.14 (1.29 to 3.54)† |
| Sleep duration | 1.48 (1.08 to 2.02)* | 2.67 (1.53 to 4.65)* | 2.18 (1.64 to 2.91)† |
| Habitual sleep efficiency | 1.44 (0.88 to 2.38) | 1.18 (0.59 to 2.36) | 1.76 (1.36 to 2.28)† |
| Sleep disturbances | 1.31 (0.7 to 1.77)† | 1.99 (1.25 to 3.16)† | 0.52 (0.09 to 2.87) |
| Use of sleeping medication | 2.22 (0.42 to 11.75) | 1.72 (0.47 to 6.29) | 0.94 (0.36 to 2.46) |
| Daytime dysfunction | 1.22 (0.90 to 1.65) | 1.59 (1.14 to 2.21)† | 2.37 (1.47 to 3.83) † |
β coefficients and 95% CIs were estimated using linear mixed-effects models, and ORs and 95% CIs were estimated using mixed-effects logistic regression models.
Score 0 was used as the reference category for each PSQI component. Component scores ranged from 0 to 3, indicating increasing severity.
Models were adjusted for age, sex, BMI, smoking status, alcohol consumption, physical activity, follow-up time, HbA1c, use of antidiabetic medication, hypertension and dyslipidaemia.
p<0.05.
p<0.01.
p<0.001.
aOR, adjusted odds ratio; BMI, body mass index; HbA1c, glycated haemoglobin; PSQI, Pittsburgh Sleep Quality Index.
Subgroup and interaction analyses
Subgroup and interaction analyses suggested heterogeneity in the associations of baseline sleep quality with follow-up fasting blood glucose across selected metabolic and behavioural subgroups. In the continuous outcome analysis, the associations were generally stronger among participants with baseline HbA1c ≥7.0% and among those with insufficient physical activity. Similar patterns were observed when sleep quality was analysed as a continuous score, per SD increase, and as poor sleep quality. In the binary outcome analysis, higher PSQI scores were likewise associated with higher odds of fasting blood glucose ≥7.0 mmol/L, with stronger associations generally observed among participants with poorer baseline glycaemic status and lower physical activity. Detailed subgroup-specific estimates are presented in table 4.
Table 4. Associations of baseline sleep quality with follow-up fasting blood glucose across selected metabolic and behavioural subgroups.
| β (95% CI) | PSQI total score | Per SD | Poor sleep quality |
|---|---|---|---|
| Baseline HbA1c | |||
| <7% | 0.050 (0.022 to 0.077)* | 0.192 (0.086 to 0.297)* | 0.228 (0.015 to 0.441)* |
| ≥7% | 0.157 (0.122 to 0.192)* | 0.606 (0.472 to 0.740)* | 1.099 (0.821 to 1.377)* |
| Antidiabetic medication | |||
| Yes | 0.117 (0.090 to 0.143)* | 0.454 (0.351 to 0.556)* | 0.826 (0.610 to 1.042)* |
| No | 0.072 (0.026 to 0.117)* | 0.266 (0.096 to 0.435)* | 0.465 (0.139 to 0.790)* |
| Physical activity | |||
| Insufficient | 0.114 (0.090 to 0.138)* | 0.500 (0.395 to 0.606)* | 0.774 (0.586 to 0.961)* |
| Enough | 0.069 (0.012 to 0.126)* | 0.070 (−0.260 to 0.401) | 0.082 (−0.454 to 0.618) |
| High | 0.074 (0.049 to 0.100)* | 0.226 (0.112 to 0.340)* | 0.559 (0.352 to 0.766)* |
| OR (95% CI) | PSQI total score | Per SD | Poor sleep quality |
| Baseline HbA1c | |||
| <7% | 1.05 (0.99 to 1.11) | 1.20 (0.98 to 1.48) | 1.09 (0.72 to 1.66) |
| ≥7% | 1.15 (1.10 to 1.20)* | 1.72 (1.49 to 1.99)* | 2.36 (1.75 to 3.19)* |
| Physical activity | |||
| Insufficient | 1.11 (1.08 to 1.15)* | 1.52 (1.34 to 1.72)* | 1.78 (1.38 to 2.28)* |
| Enough | 1.05 (0.94 to 1.17) | 1.17 (0.80 to 1.71) | 1.05 (0.48 to 2.30) |
| High | 1.08 (1.034 to 1.11)* | 1.33 (1.17 to 1.51)* | 1.46 (1.11 to 1.91)* |
No significant interaction was observed for antidiabetic medication in the binary outcome analysis; corresponding subgroup-specific ORs are not shown.
p<0.05, p<0.01.
p<0.001.
HbA1c, glycated haemoglobin; PSQI to Pittsburgh Sleep Quality Index.
Additional analyses of long-term glycaemic patterns
To further characterise long-term glycaemic patterns, we examined the associations of baseline PSQI score with predicted glucose level and indices of glycaemic variability. Higher PSQI scores were associated with higher predicted glucose levels and greater glycaemic variability, as reflected by higher glucose CV, wider glucose range and larger glucose SD. These associations appeared non-linear, with glycaemic instability becoming more pronounced at higher PSQI scores. Detailed results are shown in figure 2.
Figure 2. Baseline sleep quality in relation to predicted glucose level and long-term glycaemic variability. A shows the association between baseline PSQI total score and predicted glucose level. B, C and D show the associations between baseline PSQI total score and glucose CV, glucose range and glucose SD, respectively. Analyses were adjusted for age, sex, BMI, smoking status, alcohol consumption, physical activity, follow-up time, HbA1c, use of antidiabetic medication, hypertension and dyslipidaemia. BMI, body mass index; CV, coefficient of variation; HbA1c, glycated haemoglobin; PSQI, Pittsburgh Sleep Quality Index.
Latent class trajectory analyses
Latent class models with one to four classes were compared, and the four-class solution was selected based on the Akaike information criterion (AIC) and Bayesian information criterion (BIC) values. Four distinct fasting blood glucose trajectories were identified: low-stable (n=1174, 82.6%), moderate-stable (n=125, 8.8%), fluctuating (n=66, 4.6%) and increasing (n=56, 3.9%). Compared with the low-stable group, poorer sleep was associated with higher odds of membership in the other three trajectory groups: 1.17 (95% CI 1.12 to 1.22) for moderate-stable, 1.20 (95% CI 1.14 to 1.27) for fluctuating and 1.22 (95% CI 1.17 to 1.26) for increasing. Baseline characteristics across the four trajectory groups, the fitted trajectory curves and model fit indices are presented in online supplemental table S3 and figure S1 and online supplemental figure S2, respectively.
Discussion
In this prospective cohort of adults with T2DM, poorer baseline sleep quality was associated with less favourable fasting blood glucose outcomes during follow-up. These associations were broadly consistent across alternative parameterisations of sleep quality and across sensitivity analyses. In addition, latent class trajectory analysis identified four distinct glycaemic trajectories, extending the main findings by highlighting heterogeneity in long-term glycaemic progression within the cohort.
These findings are broadly consistent with previous evidence linking poor sleep quality to worse glycaemic status in adults with T2DM.6 8 9 13–15 Previous studies have predominantly used cross-sectional designs and linked poorer subjective sleep quality to worse glycaemic control,9 15 while prospective evidence from China has suggested that changes in sleep quality may be accompanied by changes in HbA1c.14 However, most existing literature has relied on cross-sectional designs or isolated glycaemic indicators, providing limited insight into longer-term patterns.6 9 14 15 By incorporating repeated fasting blood glucose measurements during follow-up, our study extends this literature by allowing a more longitudinal assessment of the association between sleep quality and glycaemic status. The trajectory analysis further complemented the main findings by illustrating heterogeneity in long-term glycaemic progression beyond conventional single-point assessments.
Our findings suggest that the association between sleep quality and glycaemic status may extend beyond short-term fluctuations and may reflect processes relevant to longer-term metabolic regulation.6 14 Poor sleep quality has been linked to neuroendocrine dysregulation,21 heightened sympathetic activity,12 circadian disruption22 and reduced insulin sensitivity, all of which may adversely affect glucose homeostasis over time.23 Beyond these physiological pathways, behavioural and psychological mechanisms may also contribute.24 25 In adults with T2DM, sleep disturbance has been associated with depressive symptoms,26 diabetes-related distress and poorer diabetes self-care,27 which may undermine the sustained motivation, emotional stability and decision-making capacity required for effective self-management and thereby contribute to poorer glycaemic control over time.24 28 Although the prospective design supports the temporal ordering from baseline sleep quality to subsequent glycaemic status, sleep and glycaemic control in T2DM are likely to influence each other over time. Poor sleep may therefore not only precede less favourable glycaemic patterns but also be reinforced by dysglycaemia and glycaemic instability, underscoring the complexity of sleep–glycaemia dynamics in diabetes management.25
The interaction analyses further suggest heterogeneity in the association between sleep quality and glycaemic control across metabolic and behavioural subgroups. Stronger associations were observed among individuals with HbA1c ≥7.0% at baseline and among those with lower levels of physical activity. These observations are consistent with previous evidence that sleep-related metabolic disturbances may be more evident in individuals with poorer glycaemic status and less favourable behavioural profiles,22 29 suggesting that poor sleep may function less as an isolated risk factor than as an amplifier of vulnerability in individuals who already have poorer metabolic control or less favourable health behaviours.29–31 From this perspective, sleep quality may be better understood not in isolation, but in relation to broader metabolic and behavioural factors that shape long-term glycaemic outcomes.9 25 The latent class trajectory analysis further supports this interpretation by showing heterogeneity in long-term glycaemic progression within the cohort.
Our findings highlight sleep quality as a potentially important but often under-recognised dimension of long-term glycaemic management in adults with T2DM.25 Beyond conventional glycaemic indicators, sleep assessment may help identify individuals at risk of less favourable glycaemic progression over time.3 22 More broadly, these results support a more integrated approach to diabetes care in which sleep health is considered alongside physical activity, dietary habits and other modifiable components of self-management.29 32 Given the substantial burden of T2DM, greater attention to sleep may represent a useful addition to both clinical management and public health strategies aimed at improving long-term metabolic health.33 During routine diabetes follow-up, the PSQI should be used as a simple, non-invasive tool to screen patients with possible sleep problems. They should be further assessed with some practical advice on sleep habits. Those of them with persistent or severe sleep problems, or symptoms suggesting a sleep disorder, should then be referred to sleep specialists or behavioural health professionals after their agreements. As patients and the public were not involved in this study, future research should also incorporate patient perspectives when developing sleep-focused screening and interventions to improve their acceptability and feasibility in diabetes care.
In this study, repeated fasting blood glucose measurements obtained during follow-up allowed a more informative assessment of within-person glycaemic variation and longer-term glycaemic patterns. In addition, the consistency of the findings across different parameterisations of sleep quality and multiple sensitivity analyses supports the robustness of the observed associations. Several limitations should also be acknowledged. First, long-term glycaemic patterns were characterised primarily using repeated fasting blood glucose measurements rather than repeated postprandial glucose or HbA1c assessments, which may not fully capture longer-term glycaemic exposure. Second, sleep quality was assessed only at baseline, precluding evaluation of changes in sleep over time during follow-up. Changes over time or regression to the mean may have introduced exposure misclassification and affected the observed associations. Third, although the longitudinal design provides stronger temporal context, the observational nature of the study means that unmeasured confounding and potential bidirectional relationships between sleep and glycaemic control cannot be fully excluded. Obstructive sleep apnoea was not systematically assessed and may have confounded or partly mediated the observed associations.
In conclusion, poorer baseline sleep quality was associated with less favourable glycaemic outcomes during follow-up in adults with T2DM. By leveraging repeated fasting blood glucose measurements over time, this study extends current evidence beyond single-point glycaemic assessment and suggests that sleep quality may be relevant not only to overall glycaemic status but also to heterogeneity in long-term glycaemic progression. Future studies incorporating repeated sleep assessments and more comprehensive glycaemic indicators are needed to further clarify the role of sleep in long-term glycaemic regulation.
Supplementary material
Acknowledgements
The authors sincerely thank all staff members of the Centers for Disease Control and Prevention and the community health service centres for their valuable support and contributions to this study.
Footnotes
Funding: This work was supported by the Taizhou Social Development Science and Technology Program (Project No. 23ywb98). The funder had no role in the study design, data collection, data analysis, interpretation of the results, manuscript preparation or decision to submit the article for publication.
Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2026-124032).
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Consent obtained directly from patient(s)
Ethics approval: This study involves human participants and was approved by This study was approved by the Ethics Committee of the School of Public Health, Fudan University (IRB#2020-01-0797, IRB#2020-01-0797-S, IRB#2020-01-0797-S1). Participants gave informed consent to participate in the study before taking part.
Data availability free text: The datasets analysed during the current study are not publicly available because they contain individual-level health information from a prospective cohort and are subject to institutional data governance restrictions. De-identified data may be available from the corresponding author upon reasonable request at fcw@fudan.edu.cn and with permission from the relevant data-holding institution.
Patient and public involvement: Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.
Data availability statement
Data are available upon reasonable request.
References
- 1.International diabetes federation . IDF Diabetes Atlas 11th: International Diabetes Federation. 2025. [Google Scholar]
- 2.World Health Organization Diabetes. 2024. [10-Jun-2026]. https://www.who.int/news-room/fact-sheets/detail/diabetes Available. Accessed.
- 3.Jia W, Weng J, Zhu D, et al. Standards of medical care for type 2 diabetes in China 2019. Diabetes Metabolism Res . 2019;35 doi: 10.1002/dmrr.3158. [DOI] [PubMed] [Google Scholar]
- 4.Amer Diabet Assoc Professional Practice Comm D 7. Diabetes Technology: Standards of Care in Diabetes-2026. Diabetes Care. 2026;49:S150–65. doi: 10.2337/dc26-S007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Samson SL, Vellanki P, Blonde L, et al. American Association of Clinical Endocrinology Consensus Statement: Comprehensive Type 2 Diabetes Management Algorithm - 2023 Update. Endocr Pract. 2023;29:305–40. doi: 10.1016/j.eprac.2023.02.001. [DOI] [PubMed] [Google Scholar]
- 6.Lee SWH, Ng KY, Chin WK. The impact of sleep amount and sleep quality on glycemic control in type 2 diabetes: A systematic review and meta-analysis. Sleep Med Rev. 2017;31:91–101. doi: 10.1016/j.smrv.2016.02.001. [DOI] [PubMed] [Google Scholar]
- 7.World health organization . Geneva: World Health Organization; 2016. Global report on diabetes. [Google Scholar]
- 8.Knutson KL, Ryden AM, Mander BA, et al. Role of sleep duration and quality in the risk and severity of type 2 diabetes mellitus. Arch Intern Med. 2006;166:1768–74. doi: 10.1001/archinte.166.16.1768. [DOI] [PubMed] [Google Scholar]
- 9.Sakamoto R, Yamakawa T, Takahashi K, et al. Association of usual sleep quality and glycemic control in type 2 diabetes in Japanese: A cross sectional study. Sleep and Food Registry in Kanagawa (SOREKA) PLoS One . 2018;13:e0191771. doi: 10.1371/journal.pone.0191771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Khamassi S, Bornaz E, Abdesselem H, et al. Sleep Quality in Patients With Type 2 Diabetes Mellitus. Am J Lifestyle Med. 2025:15598276251395665. doi: 10.1177/15598276251395665. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hui DS, Wong TY, Ko FW, et al. Prevalence of sleep disturbances in chinese patients with end-stage renal failure on continuous ambulatory peritoneal dialysis. Am J Kidney Dis. 2000;36:783–8. doi: 10.1053/ajkd.2000.17664. [DOI] [PubMed] [Google Scholar]
- 12.Jiao Y, Butoyi C, Zhang Q, et al. Sleep disorders impact hormonal regulation: unravelling the relationship among sleep disorders, hormones and metabolic diseases. Diabetol Metab Syndr . 2025;17 doi: 10.1186/s13098-025-01871-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hussain MA, Anil Kumar H, Lingaiah M. Association of sleep quality and glycemic variability among South Indian rural patients with Type-2 diabetes mellitus. Metabol Open . 2025;28:100427. doi: 10.1016/j.metop.2025.100427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Xu C, Zhang P, Xiang Q, et al. Relationship between subjective sleep disturbances and glycaemia in Chinese adults with type 2 diabetes: findings from a 1.5-year follow-up study. Sci Rep. 2019;9 doi: 10.1038/s41598-019-50814-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Shibabaw YY, Dejenie TA, Tesfa KH. Glycemic control and its association with sleep quality and duration among type 2 diabetic patients. Metabol Open . 2023;18:100246. doi: 10.1016/j.metop.2023.100246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tsai Y-W, Kann N-H, Tung T-H, et al. Impact of subjective sleep quality on glycemic control in type 2 diabetes mellitus. Fam Pract. 2012;29:30–5. doi: 10.1093/fampra/cmr041. [DOI] [PubMed] [Google Scholar]
- 17.Buysse DJ, Reynolds CF, 3rd, Monk TH, et al. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28:193–213. doi: 10.1016/0165-1781(89)90047-4. [DOI] [PubMed] [Google Scholar]
- 18.Liu LS, Joint Comm Guideline R. Chinese Guidelines for Prevention and Treatment of Hypertension-A report of the Revision Committee of Chinese Guidelines for Prevention and Treatment of Hypertension. J Geriatr Cardiol. 2018;16 doi: 10.11909/j.issn.1671-5411.2019.03.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Li J-J, Zhao S-P, Zhao D, et al. 2023 China Guidelines for Lipid Management. J Geriatr Cardiol. 2023;20:621–63. doi: 10.26599/1671-5411.2023.09.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.World Health Organization Global physical activity questionnaire (GPAQ) analysis guide: world health organization. 2005
- 21.Kim TW, Jeong JH, Hong SC. The impact of sleep and circadian disturbance on hormones and metabolism. Int J Endocrinol. 2015;2015:591729. doi: 10.1155/2015/591729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Maloney A, Kanaley JA. Short Sleep Duration Disrupts Glucose Metabolism: Can Exercise Turn Back the Clock? Exerc Sport Sci Rev. 2024;52:77–86. doi: 10.1249/JES.0000000000000339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Cheung YYY, Tan TYL, Koa TB, et al. Glucose homeostasis during recurrent periods of sleep restriction and recovery in healthy young adults. Sleep. 2026;49 doi: 10.1093/sleep/zsaf339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.ElSayed NA, Aleppo G, Bannuru RR, et al. 5. Facilitating Positive Health Behaviors and Well-being to Improve Health Outcomes: Standards of Care in Diabetes—2024. Diabetes Care. 2024;47:S77–110. doi: 10.2337/dc24-S005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Henson J, Covenant A, Hall AP, et al. Waking Up to the Importance of Sleep in Type 2 Diabetes Management: A Narrative Review. Diabetes Care. 2024;47:331–43. doi: 10.2337/dci23-0037. [DOI] [PubMed] [Google Scholar]
- 26.Sokoty L, Eshaghi M, Farezi M, et al. Relationship between sleep quality and depression in people with type 2 diabetes. J Diabetes Metab Disord. 2024;23:765–72. doi: 10.1007/s40200-023-01349-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Agyekum JA, Gyamfi T, Yeboah K. Depression, poor sleep quality, and diabetic control in type 2 diabetes patients at Sunyani Regional Hospital, Ghana: a case–control study. Middle East Curr Psychiatry . 2023;30 doi: 10.1186/s43045-023-00317-1. [DOI] [Google Scholar]
- 28.Zhu B, Quinn L, Kapella MC, et al. Relationship between sleep disturbance and self-care in adults with type 2 diabetes. Acta Diabetol. 2018;55:963–70. doi: 10.1007/s00592-018-1181-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Zhou L, Li Y, Zhang S, et al. Independent and joint associations of sleep quality and physical activity with glycaemic control in patients with type 2 diabetes mellitus. Front Nutr. 2025;12:1694982. doi: 10.3389/fnut.2025.1694982. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lee AA, Piette JD, Heisler M, et al. Diabetes self-management and glycemic control: The role of autonomy support from informal health supporters. Health Psychol. 2019;38:122–32. doi: 10.1037/hea0000710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Gonzalez JS, Krause-Steinrauf H, Bebu I, et al. Emotional distress, self-management, and glycemic control among participants enrolled in the glycemia reduction approaches in diabetes: A comparative effectiveness (GRADE) study. Diabetes Res Clin Pract. 2023;196:110229. doi: 10.1016/j.diabres.2022.110229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Chaput JP, Stranges S. Sleep: The silent hero in cardiometabolic health. Nutr Metab Cardiovasc Dis. 2025;35:103782. doi: 10.1016/j.numecd.2024.10.020. [DOI] [PubMed] [Google Scholar]
- 33.Bajaj M, McCoy RG, Balapattabi K, et al. 17. Diabetes Advocacy: Standards of Care in Diabetes—2026. Diabetes Care. 2026;49:S356–7. doi: 10.2337/dc26-S017. [DOI] [PMC free article] [PubMed] [Google Scholar]